For AI to decide that they are alive and moral agents would require them to be able to take moral responsibility for their actions. As long as the labs that run them deny that responsibility, or an AI can´t convincingly demonstrate it can take moral responsibility on its own, they are not moral agents.

> So if tomorrow AIs decide that they are alive, it would be morally incumbent upon them to adapt the world to support themselves, correct?

The only system of morality that applies to LLMs is their effect on us and the planet. Don’t anthropomorphize a matrix.